Machine learning model inference endpoint that uses the webcam
This example creates a web endpoint that uses a Huggingface model for object detection.
The web endpoint takes an image from their webcam, and sends it to a Modal web endpoint. The Modal web endpoint in turn calls a Modal function that runs the actual model.
If you run this, it will look something like this:
Live demo
Take a look at the deployed app.
A couple of caveats:
- This is not optimized for latency: every prediction takes about 1s, and there’s an additional overhead on the first prediction since the containers have to be started and the model initialized.
- This doesn’t work on iPhone unfortunately due to some issues with HTML5 webcam components
Code
Starting with imports:
import base64
import io
from pathlib import Path
from fastapi import FastAPI, Request, Response
from fastapi.staticfiles import StaticFiles
from modal import (
Image,
Mount,
Stub,
asgi_app,
method,
)We need to install transformers which is a package Huggingface uses for all their models, but also Pillow which lets us work with images from Python, and a system font for drawing.
This example uses the facebook/detr-resnet-50 pre-trained model, which is downloaded
once at image build time using the download_model function and saved into the image.
‘Baking’ models into the modal.Image at build time provided the fastest cold start.
model_repo_id = "facebook/detr-resnet-50"
def download_model():
from huggingface_hub import snapshot_download
snapshot_download(repo_id=model_repo_id, cache_dir="/cache")
stub = Stub("example-webcam-object-detection")
image = (
Image.debian_slim()
.pip_install(
"huggingface-hub==0.16.4",
"Pillow",
"timm",
"transformers",
)
.apt_install("fonts-freefont-ttf")
.run_function(download_model)
)Prediction function
The object detection function has a few different features worth mentioning:
- There’s a container initialization step in the
__enter__method, which runs on every container start. This lets us load the model only once per container, so that it’s reused for subsequent function calls. - Above we stored the model in the container image. This lets us download the model only when the image is (re)built, and not everytime the function is called.
- We’re running it on multiple CPUs for extra performance
Note that the function takes an image and returns a new image. The input image is from the webcam The output image is an image with all the bounding boxes and labels on them, with an alpha channel so that most of the image is transparent so that the web interface can render it on top of the webcam view.
@stub.cls(
cpu=4,
image=image,
)
class ObjectDetection:
def __enter__(self):
from transformers import DetrForObjectDetection, DetrImageProcessor
self.feature_extractor = DetrImageProcessor.from_pretrained(
model_repo_id,
cache_dir="/cache",
)
self.model = DetrForObjectDetection.from_pretrained(
model_repo_id,
cache_dir="/cache",
)
@method()
def detect(self, img_data_in):
# Based on https://huggingface.co/spaces/nateraw/detr-object-detection/blob/main/app.py
from PIL import Image, ImageColor, ImageDraw, ImageFont
# Read png from input
image = Image.open(io.BytesIO(img_data_in)).convert("RGB")
# Make prediction
import torch
inputs = self.feature_extractor(image, return_tensors="pt")
outputs = self.model(**inputs)
img_size = torch.tensor([tuple(reversed(image.size))])
processed_outputs = (
self.feature_extractor.post_process_object_detection(
outputs=outputs,
target_sizes=img_size,
threshold=0,
)
)
output_dict = processed_outputs[0]
# Grab boxes
keep = output_dict["scores"] > 0.7
boxes = output_dict["boxes"][keep].tolist()
scores = output_dict["scores"][keep].tolist()
labels = output_dict["labels"][keep].tolist()
# Plot bounding boxes
colors = list(ImageColor.colormap.values())
font = ImageFont.truetype(
"/usr/share/fonts/truetype/freefont/FreeMono.ttf", 18
)
output_image = Image.new("RGBA", (image.width, image.height))
output_image_draw = ImageDraw.Draw(output_image)
for _score, box, label in zip(scores, boxes, labels):
color = colors[label % len(colors)]
text = self.model.config.id2label[label]
box = tuple(map(int, box))
output_image_draw.rectangle(box, outline=color)
output_image_draw.text(
box[:2], text, font=font, fill=color, width=3
)
# Return PNG as bytes
with io.BytesIO() as output_buf:
output_image.save(output_buf, format="PNG")
return output_buf.getvalue()Defining the web interface
To keep things clean, we define the web endpoints separate from the prediction function. This will introduce a tiny bit of extra latency (every web request triggers a Modal function call which will call another Modal function) but in practice the overhead is much smaller than the overhead of running the prediction function etc.
We also serve a static html page which contains some tiny bit of Javascript to capture the webcam feed and send it to Modal.
web_app = FastAPI()
static_path = Path(__file__).with_name("webcam").resolve()The endpoint for the prediction function takes an image as a data URI and returns another image, also as a data URI:
@web_app.post("/predict")
async def predict(request: Request):
# Takes a webcam image as a datauri, returns a bounding box image as a datauri
body = await request.body()
img_data_in = base64.b64decode(body.split(b",")[1]) # read data-uri
img_data_out = ObjectDetection().detect.remote(img_data_in)
output_data = b"data:image/png;base64," + base64.b64encode(img_data_out)
return Response(content=output_data)Exposing the web server
Let’s take the Fast API app and expose it to Modal.
@stub.function(
mounts=[Mount.from_local_dir(static_path, remote_path="/assets")],
)
@asgi_app()
def fastapi_app():
web_app.mount("/", StaticFiles(directory="/assets", html=True))
return web_appRunning this locally
You can run this as an ephemeral app, by running
modal serve webcam.py